Flaviu Radulescu founded Runware in 2023 after encountering a common frustration while testing a text-to-image company. He recognized that while generative AI technology was powerful, it was often slow at producing images. To solve this, Radulescu partnered with Ioana Hreninciuc to launch Runware, a developer tool platform specializing in real-time generation of images, videos, and audio.
Since its launch, the company has experienced significant growth. It has powered more than 5 billion creations for over 100,000 developers. The product allows developers to integrate Runware’s API into their applications and generate media assets through a single interface. This eliminates the need for them to set up new infrastructure or maintain separate integrations. The platform features a custom AI inference infrastructure for open-source models, provides day-zero access to new models, and offers competitive pricing.
The company recently announced a 50 million dollar Series A funding round led by Dawn Capital. Dawn Capital Partner Shamillah Bankiya will join the board. Other participants in the round include Insight Partners and a16z Speedrun. To date, Runware has raised 66 million dollars in total funding.
The company maintains its competitiveness through cost-effective pricing and a fully unified API. This is enabled by its Sonic Inference Engine, which runs on custom AI hardware. Runware also partners with third-party AI cloud providers to automatically reroute workloads when additional memory is required. On the software side, heavy optimization of model loading and offloading allows the platform to support over 400,000 models, making any of them available for inference in real time.
The market for developer tools focused on image and video generation has attracted considerable venture capital interest. Runware views its competitors as general open-source model makers like Hugging Face, startups such as Replicate, and other generative AI model hosting services. Unlike many competitors that sell based on GPU compute time, Runware employs a model similar to Stable Diffusion and Flux, charging a cost per image generated. This allows customers to pay for what they need rather than purchasing blocks of compute time.
The new capital will be used to expand the company’s infrastructure. A key goal is to use the Sonic Inference Engine to power over 2 million models, with the larger ambition of becoming the universal API for all generative AI models. The company is also expanding rapidly into new modalities and plans to grow its current team of approximately 25 people to support these efforts.
Ultimately, Runware aims to enable applications to scale to millions of users while maintaining healthy margins. The company believes its approach makes the market more affordable, benefiting everyone from app builders to end users and helping to put powerful AI tools into more hands around the globe.